]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 15:12:07 +0000 (15:12 +0000)]
MFNetBSD:
    revision 1.118
    date: 2001/12/27 18:48:28;  author: augustss;  state: Exp;  lines: +24 -5
    Add some DIAGNOSTIC stuf that I forgot.  From Nate Williams.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 15:08:39 +0000 (15:08 +0000)]
MFNetBSD:
    revision 1.117
    date: 2001/12/27 11:27:11;  author: augustss;  state: Exp;  lines: +5 -3
    Update a comment.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 15:07:23 +0000 (15:07 +0000)]
MFNetBSD:
    revision 1.116
    date: 2001/11/21 08:18:40;  author: augustss;  state: Exp;  lines: +30 -3
    Pay more attention to if the HC is being unplugged.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 15:01:01 +0000 (15:01 +0000)]
MFNetBSD:
    revision 1.27
    date: 2001/02/21 10:19:30;  author: minoura;  state: Exp;  lines: +2 -3
    There is no reason to exclude FreeBSD :).

22 years agoFixed the nasty bug where .depend file that exists somewhere in
Ruslan Ermilov [Sun, 7 Apr 2002 14:58:12 +0000 (14:58 +0000)]
Fixed the nasty bug where .depend file that exists somewhere in
the .PATH (but not in the ${.OBJDIR}) would result in a leak of
the ${OBJS}: ${SRCS:M*.h} dependency hint.

Spotted by: fixing the broken gnu/usr.bin/cc/cc1obj build
MFC after: 1 day

22 years agobbsize and sbsize cannot ever be trusted from the disklabel, in
Poul-Henning Kamp [Sun, 7 Apr 2002 14:57:57 +0000 (14:57 +0000)]
bbsize and sbsize cannot ever be trusted from the disklabel, in
particular as there may not be one.  Remove #if 0'ed code which might
mislead people to think otherwise.

unifdef -ULOSTDIR, fsck can make lost+found on the fly.

Sponsored by: DARPA & NAI Labs

22 years agoMFNetBSD: ohci.c (1.114), ohcivar.h (1.29)
Josef Karthauser [Sun, 7 Apr 2002 14:57:31 +0000 (14:57 +0000)]
MFNetBSD: ohci.c (1.114), ohcivar.h (1.29)

    date: 2001/11/21 02:41:18;  author: augustss;
    Use a task to perform the timeout abort so we have a process context when
    sleeping.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 14:53:52 +0000 (14:53 +0000)]
MFNetBSD:
    revision 1.113
    date: 2001/11/21 02:39:31;  author: augustss;  state: Exp;  lines: +10 -4
    Some more debug stuff.

22 years agoMFNetBSD: ohci.c (1.112), uhci.c (1.147)
Josef Karthauser [Sun, 7 Apr 2002 14:49:14 +0000 (14:49 +0000)]
MFNetBSD: ohci.c (1.112), uhci.c (1.147)

    date: 2001/11/21 02:38:35;  author: augustss;
    Cast some args to bitmask_snprintf().

22 years agoMFNetBSD: ohci.c (1.111), uhci.c (1.146)
Josef Karthauser [Sun, 7 Apr 2002 14:45:38 +0000 (14:45 +0000)]
MFNetBSD: ohci.c (1.111), uhci.c (1.146)

    date: 2001/11/20 21:12:46;  author: augustss;
    Don't bother with interrupts when being disconnected.

22 years agoMFNetBSD: ohci.c (1.101), ohcivar.h (1.26)
Josef Karthauser [Sun, 7 Apr 2002 14:40:15 +0000 (14:40 +0000)]
MFNetBSD: ohci.c (1.101), ohcivar.h (1.26)

    revision 1.101
    date: 2001/02/20 15:20:32;  author: minoura;
    Re-initialize some registers after resuming from suspend.
    Some APM BIOSes do not restore them.
    Reviewd by augustss.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 14:34:23 +0000 (14:34 +0000)]
MFNetBSD:
    revision 1.97
    date: 2000/12/31 14:29:54;  author: augustss;  state: Exp;  lines: +17 -3
    Make the controller survive suspend/resume.  Thanks to Steve Woodford
    <steve@mctavish.co.uk> for testing.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 14:31:03 +0000 (14:31 +0000)]
MFNetBSD:
    revision 1.95
    date: 2000/12/13 03:09:06;  author: augustss;  state: Exp;  lines: +19 -3
    Make the ohci driver not hang suspend/resume.  It still doesn't resume
    correctly.  From itohy@netbsd.org (ITOH Yasufumi) PR kern/11714.

22 years agoMFNetBSD: ohci.c (1.100), ohcivar.h (1.25)
Josef Karthauser [Sun, 7 Apr 2002 14:25:02 +0000 (14:25 +0000)]
MFNetBSD: ohci.c (1.100), ohcivar.h (1.25)

ohci.c
    revision 1.100
    date: 2001/01/28 16:18:09;  author: augustss;  state: Exp;  lines: +7 -2
    Put a rate limiter on the scheduling overrun message.

ohcivar.h:
    revision 1.25
    date: 2001/01/28 19:01:20;  author: augustss;  state: Exp;  lines: +4 -1
    Commit second half of rate limit change.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 14:21:32 +0000 (14:21 +0000)]
MFNetBSD:
    revision 1.58
    date: 2001/11/20 23:53:26;  author: augustss;  state: Exp;  lines: +61 -33
    Create a special kernel thread to run the usb short lived tasks (instead
    of using the device discovery threads).

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 14:12:56 +0000 (14:12 +0000)]
MFNetBSD:
    revision 1.65
    date: 2002/01/03 22:20:45;  author: augustss;  state: Exp;  lines: +9 -2
    Add a DIAGNOSTIC check.  From FreeBSD.

(Just wrap these changes, that we already had, with a #ifdef DIAGNOSTIC).

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 14:10:52 +0000 (14:10 +0000)]
MFNetBSD:
    revision 1.62
    date: 2002/01/02 20:55:58;  author: augustss;  state: Exp;  lines: +11 -10
    Some more usb_proc_ptr changes.
    Also some minor stylistic changes.

    revision 1.63
    date: 2002/01/02 20:58:12;  author: augustss;  state: Exp;  lines: +4 -4
    Fix typo in last commit.

    revision 1.64
    date: 2002/01/02 22:44:44;  author: augustss;  state: Exp;  lines: +7 -4
    Add a comment.

The usb_proc_ptr changes were introduced in 1.62 and backed out in 1.64.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 14:02:31 +0000 (14:02 +0000)]
MFNetBSD:
    revision 1.61
    date: 2001/12/31 15:55:51;  author: augustss;  state: Exp;  lines: +3 -5
    Delay bus enumeration a little in case the controller is a companion
    controller.  This way the main controller can gain ownership of the port
    before enumeration starts.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 13:59:32 +0000 (13:59 +0000)]
MFNetBSD:
    revision 1.59
    date: 2001/11/26 20:16:55;  author: augustss;  state: Exp;  lines: +3 -3
    Change wchan name for usb task thread.

22 years agoMFNetBSD: usb.c (1.58)
Josef Karthauser [Sun, 7 Apr 2002 13:58:06 +0000 (13:58 +0000)]
MFNetBSD: usb.c (1.58)

    date: 2001/11/20 23:53:26;  author: augustss;  state: Exp;  lines: +61 -33
    Create a special kernel thread to run the usb short lived tasks (instead
    of using the device discovery threads).

The change to usb_port.h is a little hack whilst I address an issue
with NetBSD.

22 years agoRegen.
Nick Hibma [Sun, 7 Apr 2002 13:47:19 +0000 (13:47 +0000)]
Regen.

22 years agoAdd Scanlogic SL11R IDE Adapter.
Nick Hibma [Sun, 7 Apr 2002 13:45:24 +0000 (13:45 +0000)]
Add Scanlogic SL11R IDE Adapter.

22 years agoSynchronise the code style with NetBSD. (No functional changes).
Josef Karthauser [Sun, 7 Apr 2002 13:37:46 +0000 (13:37 +0000)]
Synchronise the code style with NetBSD.  (No functional changes).

22 years agoAdd code to try the nmount(2) syscall when mount(2) failed with
Maxime Henrion [Sun, 7 Apr 2002 13:28:18 +0000 (13:28 +0000)]
Add code to try the nmount(2) syscall when mount(2) failed with
EOPNOTSUPP.  This will make things less painful when I will commit
the conversion of devfs, fdescfs and pseudofs to nmount.

Reviewed by: phk

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 13:27:56 +0000 (13:27 +0000)]
MFNetBSD:
    revision 1.69
    date: 2001/12/27 18:43:46;  author: augustss;  state: Exp;  lines: +3 -3
    Change some DIAGNOSTIC #defines.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 13:26:59 +0000 (13:26 +0000)]
MFNetBSD:
    revision 1.68
    date: 2001/12/24 21:36:16;  author: augustss;  state: Exp;  lines: +2 -1
    Add some more DIAGNOSTIC tests.
    Make usb_match_device() match on USB_PRODUCT_ANY.

22 years agoo Change kernel_vmount() interface to be more convenient : pass two
Maxime Henrion [Sun, 7 Apr 2002 13:22:47 +0000 (13:22 +0000)]
o Change kernel_vmount() interface to be more convenient : pass two
  separate strings instead of passing "foo=bar".
o Don't forget to clear the VMOUNT flag on the vnode when vfs_nmount()
  fails because the fs doesn't implement VFS_NMOUNT (and in vfs_mount()
  when the fs doesn't implement VFS_MOUNT) ; also decrement the vfs
  refcount in the !MNT_UPDATE case.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 13:20:26 +0000 (13:20 +0000)]
MFNetBSD:
    revision 1.67
    date: 2001/11/21 13:44:47;  author: augustss;  state: Exp;  lines: +3 -1
    Wrap dump routine prototypes in #ifdef USB_DEBUG

22 years agoMFNetBSD: ukbd.c (1.52), usbdi.c (1.54), usbdi.h (1.35)
Josef Karthauser [Sun, 7 Apr 2002 13:16:18 +0000 (13:16 +0000)]
MFNetBSD: ukbd.c (1.52), usbdi.c (1.54), usbdi.h (1.35)

    date: 2000/01/16 09:37:18;  author: augustss;
    Let usbd_set_polling() work on a usbd_device_handle instead of
    a usbd_interface_handle.

22 years agoThis one needs NO_WERROR too.
Dag-Erling Smørgrav [Sun, 7 Apr 2002 12:53:58 +0000 (12:53 +0000)]
This one needs NO_WERROR too.

22 years agoMFNetBSD: usbdi.c (1.93), usbdi.h (1.59)
Josef Karthauser [Sun, 7 Apr 2002 12:42:18 +0000 (12:42 +0000)]
MFNetBSD: usbdi.c (1.93), usbdi.h (1.59)

    date: 2001/12/24 21:36:15;  author: augustss;
    Add some more DIAGNOSTIC tests.
    Make usb_match_device() match on USB_PRODUCT_ANY.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 12:38:16 +0000 (12:38 +0000)]
MFNetBSD:
    revision 1.92
    date: 2001/12/12 15:38:58;  author: augustss;  state: Exp;  lines: +7 -3
    Some more paranoia checks when DIAGNOSTIC.

22 years agoMove a function to a different place in the source so that we match
Josef Karthauser [Sun, 7 Apr 2002 12:32:55 +0000 (12:32 +0000)]
Move a function to a different place in the source so that we match
NetBSD.

22 years agoUpdate $NetBSD$ ident.
Josef Karthauser [Sun, 7 Apr 2002 12:29:55 +0000 (12:29 +0000)]
Update $NetBSD$ ident.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 12:25:55 +0000 (12:25 +0000)]
MFNetBSD:
    revision 1.88
    date: 2001/11/22 04:31:01;  author: augustss;  state: Exp;  lines: +3 -3
    Improve dump routine slightly.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 12:24:55 +0000 (12:24 +0000)]
MFNetBSD:
    revision 1.87
    date: 2001/11/21 15:37:41;  author: augustss;  state: Exp;  lines: +5 -5
    Fix indentation and typos.

22 years agoWe not long need to use USBD_NO_TSLEEP.
Josef Karthauser [Sun, 7 Apr 2002 12:19:50 +0000 (12:19 +0000)]
We not long need to use USBD_NO_TSLEEP.

22 years agoMFNetBSD: usbdi.c (1.80), usbdi.h (1.51)
Josef Karthauser [Sun, 7 Apr 2002 12:18:52 +0000 (12:18 +0000)]
MFNetBSD: usbdi.c (1.80), usbdi.h (1.51)

    date: 2001/04/13 11:19:58;  author: augustss;
    Finally get rid of the UGLY and EVIL hack for avoiding tsleep().

22 years agoGCC 3.1 cleanup - add a break after default: at the end of a switch
Murray Stokely [Sun, 7 Apr 2002 12:05:05 +0000 (12:05 +0000)]
GCC 3.1 cleanup - add a break after default: at the end of a switch
statement.

22 years agoMFNetBSD: if_aue.c (1.26), if_auereg.h (1.6), if_cue.c (1.10),
Josef Karthauser [Sun, 7 Apr 2002 12:04:02 +0000 (12:04 +0000)]
MFNetBSD: if_aue.c (1.26), if_auereg.h (1.6), if_cue.c (1.10),
  if_cuereg.h (1.5), if_kue.c (1.11), if_kuereg.h (1.5)

    date: 2000/03/02 12:37:51;  author: augustss;
    Use ratecheck() to limit error messages on disconnect.
    Break out some common functionality.

22 years agoMFNetBSD: usbdi.c (1.63), usbdi.h (1.41)
Josef Karthauser [Sun, 7 Apr 2002 11:56:19 +0000 (11:56 +0000)]
MFNetBSD: usbdi.c (1.63), usbdi.h (1.41)
    date: 2000/03/02 12:37:51;  author: augustss;
    Use ratecheck() to limit error messages on disconnect.
    Break out some common functionality.

We don't have ratecheck() in our kernel yet, so I've hardcoded
usb_ratecheck() to return true always.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 11:42:39 +0000 (11:42 +0000)]
MFNetBSD:
    revision 1.110
    date: 2001/11/20 16:08:10;  author: augustss;  state: Exp;  lines: +5 -4
    Use longer reset for root hubs (as told in the spec).

22 years agoMFNetBSD: ohci.c (1.104), ohcivar.h (1.28)
Josef Karthauser [Sun, 7 Apr 2002 11:33:55 +0000 (11:33 +0000)]
MFNetBSD: ohci.c (1.104), ohcivar.h (1.28)
    date: 2001/09/28 23:57:21;  author: augustss;
    Reenable RHSC interrupt after one second so hot plugging works.
    (From OpenBSD.)

22 years agoReduce the difference between our version and NetBSD's. (No
Josef Karthauser [Sun, 7 Apr 2002 11:29:31 +0000 (11:29 +0000)]
Reduce the difference between our version and NetBSD's.  (No
functional changes.)

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 11:20:41 +0000 (11:20 +0000)]
MFNetBSD:
    revision 1.57
    date: 2001/11/20 16:08:37;  author: augustss;  state: Exp;  lines: +3 -2
    Add a comment.

22 years agoMFNetBSD: ohci.c (1.109), uhci.c (1.144), uhub.c (1.56), usb.c (1.57),
Josef Karthauser [Sun, 7 Apr 2002 11:19:05 +0000 (11:19 +0000)]
MFNetBSD: ohci.c (1.109), uhci.c (1.144), uhub.c (1.56), usb.c (1.57),
  usbdi.c (1.86), usbdivar.h (1.66)
  [Some partial, because most of this was merged in a while ago]

    date: 2001/11/20 13:48:03;  author: augustss;
    Keep track of device speed for USB 2.0.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 10:59:02 +0000 (10:59 +0000)]
MFNetBSD:
    revision 1.55
    date: 2001/11/16 02:21:54;  author: augustss;  state: Exp;  lines: +4 -4
    Better error message.

22 years agoMFNetBSD: uhub.c (1.54), usb_subr.c (1.92)
Josef Karthauser [Sun, 7 Apr 2002 10:57:42 +0000 (10:57 +0000)]
MFNetBSD: uhub.c (1.54), usb_subr.c (1.92)
    date: 2001/11/16 01:57:47;  author: augustss;
    Handle devices that disappear during reset gracefully.

22 years agoGet us to WARNS=2 with a bit of constification and some printf fiddling.
Poul-Henning Kamp [Sun, 7 Apr 2002 10:57:00 +0000 (10:57 +0000)]
Get us to WARNS=2 with a bit of constification and some printf fiddling.

Pick up the c-partitions magicness from sys/disklabel.h instead
of defining our own magicness for it, remove trivial comment.

Sponsored by: DARPA and NAI Labs.

22 years agoMFNetBSD: usbdi.c (1.85), usbdi.h (1.54), usbdi_util.c (1.37)
Josef Karthauser [Sun, 7 Apr 2002 10:50:41 +0000 (10:50 +0000)]
MFNetBSD: usbdi.c (1.85), usbdi.h (1.54), usbdi_util.c (1.37)
    date: 2001/11/15 15:15:59;  author: augustss;
    usbd_interface2device_handle() cannot fail.

22 years agoMFNetBSD: usbdi.c (1.82), usbdivar.h (1.64)
Josef Karthauser [Sun, 7 Apr 2002 10:41:51 +0000 (10:41 +0000)]
MFNetBSD: usbdi.c (1.82), usbdivar.h (1.64)
    date: 2001/11/10 16:54:56;  author: augustss;
    Add some dump routines for debugging.

22 years agoAdd comment to supplement my last commit.
Murray Stokely [Sun, 7 Apr 2002 10:40:31 +0000 (10:40 +0000)]
Add comment to supplement my last commit.

Requested by:  obrien

22 years agoNuke a totally pointless optional debug option
Poul-Henning Kamp [Sun, 7 Apr 2002 10:39:23 +0000 (10:39 +0000)]
Nuke a totally pointless optional debug option

Sponsored by: DARPA and NAI Labs.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 10:38:07 +0000 (10:38 +0000)]
MFNetBSD:
    revision 1.81
    date: 2001/04/17 00:05:33;  author: augustss;  state: Exp;  lines: +7 -1
    Add a DIAGNOSTIC.

22 years agoMFNetBSD: usb_subr.c (1.90), usbdi.c (1.83), usbdivar.h (1.65)
Josef Karthauser [Sun, 7 Apr 2002 10:36:15 +0000 (10:36 +0000)]
MFNetBSD: usb_subr.c (1.90), usbdi.c (1.83), usbdivar.h (1.65)
    date: 2001/11/10 17:11:38;  author: augustss;  state: Exp;  lines: +1 -2
    Get rid of unused abort_handle.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 10:31:39 +0000 (10:31 +0000)]
MFNetBSD:
    revision 1.52
    date: 2001/10/26 17:53:59;  author: augustss;  state: Exp;  lines: +8 -2
    Compare pointer with NULL instead of using them as a condition.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 10:29:48 +0000 (10:29 +0000)]
MFNetBSD:
    revision 1.51
    date: 2001/10/24 15:30:17;  author: augustss;  state: Exp;  lines: +3 -3
    Better debug message.

22 years agoGCC >= 3 and C99 handle zero-length arrays differently than older
Murray Stokely [Sun, 7 Apr 2002 10:27:24 +0000 (10:27 +0000)]
GCC >= 3 and C99 handle zero-length arrays differently than older
versions of GCC.  With this change, sysinstall compiles and works fine
with GCC 3.1 or 2.95.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 10:24:52 +0000 (10:24 +0000)]
MFNetBSD:
    revision 1.107
    date: 2001/11/10 17:09:28;  author: augustss;  state: Exp;  lines: +31 -26
    Improve dump routines.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 10:21:16 +0000 (10:21 +0000)]
MFNetBSD:
    revision 1.106
    date: 2001/11/09 15:01:57;  author: augustss;  state: Exp;  lines: +73 -57
    Fix a bug in xfer abort processing when the HC executes ahead of what
    the driver aborts.
    Don't block RHSC interrupts.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 10:12:40 +0000 (10:12 +0000)]
MFNetBSD:
    revision 1.105
    date: 2001/11/07 02:55:04;  author: augustss;  state: Exp;  lines: +17 -3
    Improve some debug messages.

22 years agoMFNetBSD: ohci.c (1.104), ohcireg.h (1.28)
Josef Karthauser [Sun, 7 Apr 2002 10:09:23 +0000 (10:09 +0000)]
MFNetBSD: ohci.c (1.104), ohcireg.h (1.28)
    date: 2001/09/28 23:57:21;  author: augustss;
    Reenable RHSC interrupt after one second so hot plugging works.
    (From OpenBSD.)

22 years agoRemove some last vestages of _interactiveHack.
David E. O'Brien [Sun, 7 Apr 2002 10:08:55 +0000 (10:08 +0000)]
Remove some last vestages of _interactiveHack.
Also tie stderr to something for the invocation of pkg_add.

Sponsored by: FreeBSD Mall, Inc.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 10:04:16 +0000 (10:04 +0000)]
MFNetBSD:
    revision 1.103
    date: 2001/09/11 07:00:19;  author: augustss;  state: Exp;  lines: +4 -1
    Don't go top mode OPERATIONAL (before reset) on startup even
    if BIOS claims to have initialized the controller.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 10:02:45 +0000 (10:02 +0000)]
MFNetBSD:
    revision 1.102
    date: 2001/04/01 15:00:29;  author: augustss;  state: Exp;  lines: +3 -1
    Add two missing splx() (inside DIAGNOSTIC).  From OpenBSD.

22 years agoConstify.
Poul-Henning Kamp [Sun, 7 Apr 2002 09:58:28 +0000 (09:58 +0000)]
Constify.

Sponsored by: DARPA & NAI Labs.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 09:57:23 +0000 (09:57 +0000)]
MFNetBSD:
revision 1.142
date: 2001/10/25 02:08:13;  author: augustss;  lines: +14 -16
Change reset sequence to get status change bits set right on startup.

22 years agoMFNetBSD:
Josef Karthauser [Sun, 7 Apr 2002 09:52:49 +0000 (09:52 +0000)]
MFNetBSD:
revision 1.141
date: 2001/10/24 21:04:04;  author: augustss;  Exp;  lines: +20 -4
Some more paranoia tests when entering the interrupt routine.

22 years agoFirst version of the translation of the Security section
Sebastien Gioria [Sun, 7 Apr 2002 09:04:04 +0000 (09:04 +0000)]
First version of the translation of the Security section

22 years agoRemove a comment which relates to the old name cache code, which
David Malone [Sun, 7 Apr 2002 08:58:31 +0000 (08:58 +0000)]
Remove a comment which relates to the old name cache code, which
was replaced in 1997.

Approved by: phk

22 years agoA variable had been unnecessarily assigned a bogus value because gcc was
Garance A Drosehn [Sun, 7 Apr 2002 08:12:39 +0000 (08:12 +0000)]
A variable had been unnecessarily assigned a bogus value because gcc was
"confused" about it being unassigned.  In fact, gcc was right.  Fix the
real problem by setting that variable before break-ing out of a select
statement so gcc is happy, and then remove the unnecessary assignment.

Reported by: a user wondering why lpd syslog-ed about "compiler confusion"
MFC after: 12 days

22 years agoRename a variable from 'user' to 'userid' to avoid some compiler warnings.
Garance A Drosehn [Sun, 7 Apr 2002 08:01:42 +0000 (08:01 +0000)]
Rename a variable from 'user' to 'userid' to avoid some compiler warnings.

MFC after: 12 days

22 years agoAdd yet another ATA133 Promise chip.
Søren Schmidt [Sun, 7 Apr 2002 07:53:34 +0000 (07:53 +0000)]
Add yet another ATA133 Promise chip.

22 years agoRearrange all the error returns from sendfile() to make sure the original
Garance A Drosehn [Sun, 7 Apr 2002 07:48:32 +0000 (07:48 +0000)]
Rearrange all the error returns from sendfile() to make sure the original
input file and any temporary (filter) file are closed upon return, and
that is generally done at the end of the routine.  This should make it
easier for a later update (not yet written) to implement a "resend_copies"
option.

MFC after: 12 days

22 years agoReduce the duplication of code for error handling in _aio_aqueue().
Alan Cox [Sun, 7 Apr 2002 07:17:59 +0000 (07:17 +0000)]
Reduce the duplication of code for error handling in _aio_aqueue().

22 years agoLater versions of gcc object very strongly to the -C flag being passed
Peter Wemm [Sun, 7 Apr 2002 06:28:38 +0000 (06:28 +0000)]
Later versions of gcc object very strongly to the -C flag being passed
as -CC.  It makes no difference to gcc in the tree, but this change is
necessary for gcc-3.1 to work.

22 years agoAdd net/cvsup-without-gui and print/acroread4 to the disc3 package list.
Murray Stokely [Sun, 7 Apr 2002 06:03:33 +0000 (06:03 +0000)]
Add net/cvsup-without-gui and print/acroread4 to the disc3 package list.

22 years agoRe-arrange how output filters (of=) are handled for queues going to
Garance A Drosehn [Sun, 7 Apr 2002 05:37:27 +0000 (05:37 +0000)]
Re-arrange how output filters (of=) are handled for queues going to
remote machines.  Now they really are handled *exactly* the same as
input filters (if=) for remote queues, except that they are started
with a different set of parameters.  This should fix a few subtle
bugs in output-filter processing on such queues.  It is a pretty
significant re-arranging of sendfile(), moving some of it to a new
execfilter() routine.

PR: 36552
Reviewed by: no screams from freebsd-audit
MFC after: 12 days

22 years agoWhen checking the alternate superblock, we used to copy any fields
Kirk McKusick [Sun, 7 Apr 2002 05:16:33 +0000 (05:16 +0000)]
When checking the alternate superblock, we used to copy any fields
that might have changed, then did a byte-by-byte comparison with
the alternate. If any unused fields got used, they had to be added
to the exception list. Such changes caused too many false alarms.
So, I have changed the comparison algorithm to compare a selected
set of fields that are not expected to change. This new algorithm
causes far fewer false hits and still does a good job of detecting
problems when they have really occurred. In particular, this change
should ease the transition to kernels supporting UFS2 which make
some significant changes to the superblock.

Sponsored by: DARPA, NAI Labs

22 years agoerrx()/strerror() -> err()
Chris D. Faulhaber [Sun, 7 Apr 2002 04:59:13 +0000 (04:59 +0000)]
errx()/strerror() -> err()

22 years agoFix (for the second time) kvm_getprocs() for the case where no
Dima Dorfman [Sun, 7 Apr 2002 04:47:58 +0000 (04:47 +0000)]
Fix (for the second time) kvm_getprocs() for the case where no
processes match the given criteria.  Since revision 1.60 of malloc.c,
malloc() and friends return an invalid pointer when given a size of 0.
kvm_getprocs() uses sysctl() with a NULL oldp argument to get an
initial size, but does not check whether it's 0 before passing it to
realloc() (via _kvm_realloc()).  Before the aforementioned malloc()
change, this resulted in a minimal allocation made and a valid poitner
returned, but now results in an invalid, but non-NULL, pointer being
returned.  When this is passed to sysctl(), the latter returns EFAULT
(as it should).

22 years agoTurn on NO_WERROR due to namespace pollution in krb5 headers.
Dag-Erling Smørgrav [Sun, 7 Apr 2002 04:44:16 +0000 (04:44 +0000)]
Turn on NO_WERROR due to namespace pollution in krb5 headers.

22 years agoia64 depends on ACPICA on actual hardware. It might be worth having a
Peter Wemm [Sun, 7 Apr 2002 04:36:02 +0000 (04:36 +0000)]
ia64 depends on ACPICA on actual hardware.  It might be worth having a
seperate SKI config (like we had SIMOS for alpha).

22 years agoNow that we can thorougly self compile the kernel, stop hard coding the
Peter Wemm [Sun, 7 Apr 2002 04:34:21 +0000 (04:34 +0000)]
Now that we can thorougly self compile the kernel, stop hard coding the
cross compiler names.

22 years agoWe must not let install(1) strip loader.efi when installing it, or the
Peter Wemm [Sun, 7 Apr 2002 04:27:50 +0000 (04:27 +0000)]
We must not let install(1) strip loader.efi when installing it, or the
resulting binary will be damaged and no longer work.

22 years agoSet BINDIR
Peter Wemm [Sun, 7 Apr 2002 04:26:42 +0000 (04:26 +0000)]
Set BINDIR

22 years agoAdd loader bootforth infrastructure and install it.
Peter Wemm [Sun, 7 Apr 2002 04:25:45 +0000 (04:25 +0000)]
Add loader bootforth infrastructure and install it.

22 years agoFix a relocation bug in the ia64 ld.so. Weak function pointers in shared
Peter Wemm [Sun, 7 Apr 2002 04:16:35 +0000 (04:16 +0000)]
Fix a relocation bug in the ia64 ld.so.  Weak function pointers in shared
objects were not being correctly set to zero.  Instead, the function
descriptor pointer was set to the load address of the .so object.  This
caused gcc generated binaries to segfault on exit when crtbegin.asm's
_fini code tested the __cxa_finalize() function pointer for zero.

This is a bit of a hack because of a problem nearby workaround for
find_symdef and its quirks (failures) for local symbols.  This still
needs to be fixed.

22 years agoMinor style cleanup. Remove unneeded includes, and correct or add
Dima Dorfman [Sun, 7 Apr 2002 03:59:31 +0000 (03:59 +0000)]
Minor style cleanup.  Remove unneeded includes, and correct or add
some comments.

Reviewed by: bde

22 years agoAdd warning about collating order
Andrey A. Chernov [Sun, 7 Apr 2002 03:15:51 +0000 (03:15 +0000)]
Add warning about collating order

22 years agoFix this table, it was completely out of reasonable order
Andrey A. Chernov [Sun, 7 Apr 2002 02:52:11 +0000 (02:52 +0000)]
Fix this table, it was completely out of reasonable order

22 years agoAdd -lstdc++ to the link args for these C++ tools so that they can be
Peter Wemm [Sun, 7 Apr 2002 02:40:55 +0000 (02:40 +0000)]
Add -lstdc++ to the link args for these C++ tools so that they can be
compiled with gcc-3.1.  Somebody thought it was a good idea to move
the implementation of new and delete from libgcc to libstdc++.  This
change doesn't harm the current compiler in the tree.

22 years agoFix sed(1) in the case where a last line is specified and hold space is not
Juli Mallett [Sun, 7 Apr 2002 02:29:56 +0000 (02:29 +0000)]
Fix sed(1) in the case where a last line is specified and hold space is not
specified, and then the first part of the pattern space is deleted, when
there are two or more input lines, as this results in subtraction of one from
an unsigned integral value of '0'.  That bogus value is used in one case
for a loop (that will run far too many times in this case) and a function to
search for a value within a specified range of memory, however now the range
of memory is obscenely large and a segmentation fault will occur.  This is
fixed by checking for and appropriately handling a nil pattern space as if
the specified search in memory failed, as indeed it obviously will with nil
pattern space.

Submitted by: Tim J. Robbins <tim@robbins.dropbear.id.au>
PR: bin/34813
Reviewed by: mike
MFC after: 1 day

22 years agoBandaid for LP64 machines where 'warning: cast to pointer from different
Peter Wemm [Sun, 7 Apr 2002 01:44:24 +0000 (01:44 +0000)]
Bandaid for LP64 machines where 'warning: cast to pointer from different
sized integer' bites.  The various malloc functions return pointers,
but without any prototype/declarations visible to callers, the compiler
expects them to return int.

22 years agoChange jobref and *ijoblist from int to long in order to avoid
Alan Cox [Sun, 7 Apr 2002 01:28:34 +0000 (01:28 +0000)]
Change jobref and *ijoblist from int to long in order to avoid
a catastrophe after the 2^32nd AIO operation on 64-bit architectures.

22 years agoOriginal code bugfixed. This NOT touch old sorting order a bit.
Andrey A. Chernov [Sun, 7 Apr 2002 00:49:00 +0000 (00:49 +0000)]
Original code bugfixed. This NOT touch old sorting order a bit.

1) Remove \n from COPYRIGHT macro to shut up assembler warning
2) Fix struct sigaction initialization (arguments swap)
3) Fix out of bounds negative indexing for -R arg > 127
4) Remove doubled initialization of vars already initialized statically
5) Convert ctype macros char arg to u_char to prevents sign extension

22 years agoCorrection to rev 1.1's log:
David E. O'Brien [Sat, 6 Apr 2002 23:17:36 +0000 (23:17 +0000)]
Correction to rev 1.1's log:
Break some things out of Makefile.inc that are used by several parts of
the GCC 3.1 build -- including in ../lib.

22 years agoBreak some things used by the front-ends from Makefile.inc that cannot
David E. O'Brien [Sat, 6 Apr 2002 23:16:26 +0000 (23:16 +0000)]
Break some things used by the front-ends from Makefile.inc that cannot
be used build-wide for GCC 3.1.

22 years agoBuild and install collect2. This is needed for some C++ programs.
David E. O'Brien [Sat, 6 Apr 2002 23:12:46 +0000 (23:12 +0000)]
Build and install collect2.  This is needed for some C++ programs.

22 years agoBreak some things out of Makefile.inc that cannot be used build-wide
David E. O'Brien [Sat, 6 Apr 2002 22:37:19 +0000 (22:37 +0000)]
Break some things out of Makefile.inc that cannot be used build-wide
for GCC 3.1.